Revision and Review are two distinct quality-assurance activities defined in international standards such as ISO 9000 and ISO 17100. While the terms are sometimes used interchangeably, they describe different processes with different objectives in professional translation workflows.
What is revision?
In ISO terminology, revision is a bilingual quality-control activity. It requires a systematic comparison of the source text and the target text to verify:
- accuracy of meaning
- completeness of information
- correct transfer of terminology
- fidelity to tone and intent
- absence of mistranslations
- stylistic and linguistic appropriateness
A reviser works in both languages to ensure the translation reflects the source without omission, error, or distortion.
What is review?
Review is a monolingual assessment of the final target text. It focuses on readability, naturalness, coherence, domain suitability, stylistic alignment, and audience appropriateness without systematically comparing against the source.
- readability and natural flow
- coherence and narrative consistency
- domain suitability and brand alignment
- cultural and audience appropriateness
Review validates whether the translated text is fit for purpose in its intended context.
How the two processes interact
Both processes complement each other: revision corrects linguistic errors by ensuring semantic accuracy, while review polishes style and ensures the text functions for its audience.
- Revision: ensures bilingual accuracy and fidelity to the source.
- Review: ensures stylistic quality and publication readiness.
Examples
- Legal contract: revision for clause-level accuracy; review for consistent legal tone.
- Medical leaflet: revision for correctness; review for patient-friendly language.
- Marketing brochure: light revision; intensive review for brand voice.
The balance depends on risk, text function, and client requirements.
Revision and review in AI assisted translation
When machine translation is used, revision validates MT output against the source to correct mistranslations, omissions, hallucinations, and terminology errors. Review ensures the edited output reads fluidly, maintains brand voice, and meets domain expectations.
How Trad AI supports revision and review
Trad AI strengthens both steps with document-level context, glossary enforcement, structured domain prompts, and automatic translation-memory generation. All translations run through user-owned API keys, so no text is stored or reused, aligning with ISO 9000, GDPR, and the EU AI Act while giving linguists full control over quality.
#ISOStandards #TranslationQuality #LinguisticQA #TradAI